Identifying cities that have sprawled

The exam for my class on automated analysis of spatial grids was to take North American population density grids in 1990 and 2010 and use them to identify which cities had grown spatially. The motive was not to find which pixels had experienced large increases in population; instead, it was to find out which cities had dealt with large increases in population by spreading out.

Of course, before I gave it to the students, I tried it out first.

San Diego has exploded, but Los Angeles is still about the same size

I found major urban areas and determined which of those areas grew in size. You can see the result overlaid on Google Maps by clicking on this link: US cities: sprawl [Google Maps].  In the map, red shows cities that have grown in size, while green refers to cities that have remained stable in size between 1990 and 2010. Again, this is not about population growth, but about population sprawl.  Thus, San Diego has sprawled, but Los Angeles has remained about the same size. Interestingly, New York city hasn't sprawled, but Toronto (!) has.

 If we rank the cities (I considered only cities larger than about 80 sq. kms and having an urban core of at least 30,000 residents per sq. km) by the percent growth in metropolitan area, the ranking is:

1. San Jose, CA: the king of sprawl with its high-density urban region increasing from about 300 sq. km to about 500 sq. km between 1990 and 2010
2. San Juan, Puerto Rico
3. San Diego, CA
4. Washington, DC

Interestingly, considering all that we think Canada is about, Toronto is one of the cities that has sprawled the most.  Next only to Mexico City.

But ... but, I can hear you asking. What about Phoenix? You can see city extents in 2010 according to my methodology by clicking on this link: US cities: high-density urban areas [Google Maps].  Cities like Phoenix have suburbs that are so sparse that they don't even meet my "urban" criteria of 30K residents per sq. km. Hence, they are not identified as having grown.  I used internal measures of the data to identify what "urban" means. So, if you find an official definition of suburban population density, let me know and I can rerun my analysis with the lower threshold.

You can see a longer description of my methodology, and the approaches followed by different students here: sprawl methodology [PDF file] .